WO1995019637A1 - Particle beam, in particular ionic optic reproduction system - Google Patents

Abstract

A particle beam, in particular an ionic optic reproduction system, preferably for lithographic purposes, has a particle source, in particular an ion source for reproducing on a wafer a structure designed in a masking foil as one or several transparent spots, in particular openings, through at least two electrostatic lenses arranged upstream of the wafer. One of the lenses is a so-called grating lens constituted by one or two tubular electrodes (R1, R2) and by a perforated plate arranged in the path of the beam perpendicularly to the optical axis D. The plate is formed by a masking foil M which forms the central or first electrode of the grating lens, in the direction of propagation of the beam.

Description

 Particle, in particular ion-optical imaging system

Among the various steps that must be performed to manufacture semiconductor devices, the lithography step is particularly important. Put simply, each lithography step begins with the application of a thin layer of light-sensitive (or particle beam-sensitive) material, the so-called photoresist or "resist" for short, to a wafer, in particular made of silicon. A lithography device then projects the structure present on a mask in the form of one or more transparent locations onto the wafer provided with the resist. There may be optical elements between the mask and the wafer. The extent of the projected structure of the mask on the wafer is usually much smaller than the wafer. Following the projection, the wafer is shifted and the same structure of the mask is projected onto another location on the wafer. This projecting and shifting process is repeated until the entire wafer surface is used. By subsequently developing the resist, the desired pattern is obtained on the wafer, for example in the form of resist-free areas. In further steps, the wafer can then be subjected to any of the known processing operations such as etching, ion implantation or application and diffusion of doping material. After these further steps, the wafer is checked, again covered with resist, and the entire sequence of steps mentioned above is repeated approximately 10-20 times until a checkerboard-like arrangement of identical microcircuits is finally produced on the wafer.

Most projection lithography methods used today use light to irradiate the resist, but the need for smaller structures and higher density of the components of the microcircuits has led to intensive searches for other irradiation methods which, in terms of their resolution, do not have a relatively large resolution as when using light Wavelength are limited. Great efforts have been made to use X-rays in lithography equipment, while other methods such as e.g. particle, especially ion beam lithography, some, but received much less attention.

For particle, in particular ion optical imaging systems, preferably for lithography purposes, with a particle, in particular ion source for imaging a structure located on a mask in the form of one or more openings over at least Two collecting lenses between the mask and a wafer on the wafer have already been proposed by European patent application 93 890 058.6 to design at least one of the collecting lenses as a so-called three-electrode grating lens, which consists of two tubular electrodes, between which there is a third electrode, which is called Plate with a plurality of openings, preferably designed as a grating, the plate, in particular the grating, is arranged perpendicular to the optical axis, so that the lens is divided into two areas by the plate or grating, preferably to the three electrodes of different voltages are applied. One of the lens areas of the three-electrode grating lens with a grating as the central electrode can have positive refractive power, but the second lens area can have negative refractive power, the absolute value of the refractive power of the lens area having negative refractive power (dispersing area) being less than the refractive power of the positive refractive lens area ( collecting area). With such an imaging system, the image error coefficients can largely compensate for one another, despite the different absolute amounts of the refractive powers, if the diverging area has correspondingly larger 3rd order image errors than the collecting area.

The use of a three-electrode grating lens thus makes it possible to make one image error coefficient in the imaging equations of this lens disappear, the other image error coefficients also taking small values compared to those, for example, from those described in European patent applications EP-89109553, EP -92890165, EP-92890181 known electrostatic lenses. This is achieved by introducing the plate or grid electrode, since it is only possible with such an electrode to produce an electrostatic diverging lens. In addition, in ion-optical lithography systems with three-electrode grating lenses, there is also a reduction in the resulting distortion at the point of the minimum distortion. The use of three-electrode lenses with a grid as the central electrode also allows the mask-wafer distance to be reduced compared to known systems if the distortion values are related to the same image field size. Finally, the depth of field of the imaging system is also significantly increased because the three-electrode lenses with grids as the middle electrode in each case only have very small image defects and therefore the compensation of the image defects in the lenses located between the wafer and the mask is less critical.

Essentially the same effect can be achieved if the order of diverging and converging lenses is reversed, for example if the electrode designed as a grating forms the first electrode of a diverging two-electrode grating lens, which is followed by a converging lens designed as a field lens, for example. Here, field lens or immersion lens is understood to be an arrangement of two coaxial, rotationally symmetrical electrodes (e.g. tubular) that are at different potentials.

Disturbances through the openings in the plate or through the grid openings can be reduced if the grid openings and their spacings are kept very small on the one hand (in the order of magnitude of micrometers) and on both sides of the grid, in the area of the illumination by the ion beam as similar as possible Field strengths are available.

The problem with the use of grids is that, because of the fine openings and webs, it has to be very thin, but still has to be relatively large, and is therefore very sensitive to damage, so that it is both in the course of production and then in Continuous checks are required during use in the lens. For this purpose, the grille must be removed from the machine, cleaned if necessary and returned to its place, and because of its fragility it is always exposed to the risk of damage or destruction.

The object of the invention is therefore to remedy this problem. For this purpose, the invention proposes the design of a particle, in particular ion-optical imaging system, preferably for lithography purposes, with a particle, in particular ion source for imaging a structure located on a mask film in the form of one or more transparent locations, in particular openings via at least two in the beam direction electrostatic lenses arranged in front of a wafer on the wafer, at least one of the lenses being a so-called grating lens, which is made up of one or two Tube electrode and a plate arranged in the beam path with openings is formed, the plate being arranged normal to the optical axis, before that according to the invention the plate is formed by the masking film which forms the middle or the first electrode of the grating lens in the beam direction. In this case, the mask together with the electrode following in the ion beam direction form a diverging lens, the third-order image errors of which are compensated for by the subsequent collective lens (s) except for minimal residual values.

By replacing the grating of the known design with the mask, there is only one element that requires the highest precision in manufacture and application, namely the mask itself, which can also be produced with the same precision in all other embodiments, but is omitted the further element, which is also to be manufactured with high precision and thus also errors which cause imperfections of this second element.

In a further embodiment of the invention, the ion-optical imaging system, in which the mask represents the center electrode of a three-electrode grating lens, has a converging lens in the form of a field lens or an asymmetrical single lens as a further imaging element. Here asymmetric single lens is to be understood as a three-electrode lens in which all electrodes are at different potentials. Asymmetric single lenses are used because they have considerably smaller image defects than immersion lenses with equivalent imaging properties.

In the three-electrode grating lens, whose central electrode is now the mask, the first area practically represents the illuminating lens for the mask and the second area the diverging lens (negative refractive power), the errors of which are largely compensated for by the lenses following in the beam path.

In another embodiment, in which the mask forms the first electrode of a two-electrode grating lens, a field lens follows in the beam path and then an asymmetrical single lens. In this case, the aberrations the last lens before the wafer is compensated for by the errors of the diverging lens containing the mask and the subsequent lens.

Because the mask is now placed at the location where the grating was in the known embodiment (European patent application 93 890 058.6), the grating plane now becomes the object plane. The openings in the mask form aperture lenses, as do the openings of the grating in the known design. However, since the mask is now the object to be imaged, which is imaged in the image plane, according to the optical laws, a slight change in the angle of the particles emerging from the mask openings is largely negligible in the effect on the imaging, since all emanate from an object point Particles are collected in one pixel.

In a further embodiment of the embodiment according to the invention in which the mask forms the first electrode of a two-electrode grating lens, the point of intersection of the beams ("cross-over") lies in the field-free space in the beam direction in front of the last lens in front of the wafer, and thus between the field lens and the asymmetrical single lens. In a further embodiment of the invention, it can be provided that the field lens accelerates the ions to an energy that is high enough that the so-called stochastic space charge effects (see below) are minimized. In the asymmetrical single lens, the energy of the ions can optionally be reduced to the energy desired on the wafer.

The invention is explained below, for example with reference to the drawing. 1 a, schematic and simplified in a side view, shows the structure and beam path of an ion lithography system according to the invention of the first embodiment with a crossing point, with distortion correction, FIG. 1 b shows the analog view of the first embodiment without a crossing point, FIG. 2 shows an ion lithography system of the second embodiment Distortion correction, FIGS. 3a and b also ion lithography systems of the type shown in FIGS. 1a and 1b and FIG. 2 with correction of the chromatic error, FIG. 4 in axial longitudinal section a lens having three electrodes, the middle electrode being formed by the mask, 5 likewise in axial longitudinal section a combination of a converging lens according to FIG. 4 with a further lens designed as a field lens and arranged in front of the wafer, FIG. 6, partly in axial longitudinal section and simplified, an ion projection device with a two-electrode grating lens, in 7, the course of the chromatic resolution error and the distortion as a function of the wafer position for an arrangement according to FIG. 6, FIG. 8 a view analogous to that of FIG. 1b, but with a correction lens in front of the 9 and FIG. 9 shows a planar section through a specific exemplary embodiment of an arrangement according to FIG. 6.

1 (a and b), the ion-optical imaging system has an ion source Q for imaging a structure located on a mask M onto a wafer W. The structure is present as at least one opening of the mask M. In the first embodiment, two of the electrodes, as illustrated in FIGS. 4 and 5, are designed as tube electrodes R1 and R2 and the third electrode arranged between these tube electrodes R1 and R2 is formed by the mask M. The mask M is arranged perpendicular to the optical axis D of the imaging system and divides the lens LI into two areas P and N. The three electrodes, namely the two tubular electrodes R 1 and R2 and the electrode formed by the mask M, have different potentials Ul, U2 and U3. 1 in the drawing denotes the target beam and 2 the actual beam.

In the example described, the lens area P has positive and the lens area N has negative refractive power, but the absolute amount of the refractive power of the lens area N having negative refractive power is less than the refractive power of the lens area P having positive refractive power, so that the total refractive power of the three-electrode lens G3 is positive.

From Fig. 4 and Fig. 5 it can also be seen that the tube electrode R2 of the lens area N of negative refractive power, smaller (about half as large) diameter as the part of the tube electrode Rl of the lens area P facing the mask with positive refractive power. The tension (U3 - U0) / (U2 - UO) having between the electrodes of the negative refractive power lens portion N is also clotting he ^g (about a third) as the voltage ratio (U3 - U0) / (U 1 - UO) between the electrodes of the lens region P having positive refractive power, where UO is the potential at which the kinetic energy of the charged particles used becomes zero.

On both sides of the mask M, the same field strength is sought for the area that is illuminated by the ion beam.

In the three-electrode grating lens, which now has the mask as a grating, the first area practically represents the illuminating lens for the mask and the second area the diverging lens (negative refractive power), the errors of which are largely compensated for by the subsequent converging lens L2. Because the mask is now placed at the location where the grating was in the known embodiment (European patent application 93 890 058.6), the grating plane now becomes the object plane. The openings in the mask practically form the openings of the grating, the mask openings having the same effect as the aperture lenses of the grating openings themselves. However, since, as already mentioned, the mask M is now the object to be imaged in the optical column, which is in the According to the optical laws, a slight change in the angle of the particles emerging from the mask openings in the effect on the image is negligible, since all particles originating from an object point are collected in one image point.

These explanations apply to the formation of the second lens L2 as an asymmetrical single lens as well as a field lens.

The mask M, which has the structure to be imaged in the form of openings in a film (for example made of silicon), is provided by an ion source Q with a very small virtual source size (approximately 10 μm) and this subsequent lighting device, which has two multipoles and one arranged between them Mass separator (eg ExB filter) and optionally a diagnostic system can be illuminated and located in a first embodiment between the terminal electrodes of the lens G3 consisting of three electrodes R1, R2 and M with an overall positive refractive power. In a special embodiment (Fig. La), the lens G3 creates a crossover area (Cross-over) C, i. i a real image of the ion source Q behind its focal point on the image side. The object-side focal point of the second converging lens L2 arranged in front of the wafer W is approximately at the location of the crossover C. Thus, all rays leave the converging lens L2 almost axially parallel and an approximately telecentric imaging system is obtained. This has the advantage that the imaging scale does not change with small displacements of the wafer W in the direction of the ion-optical axis.

In a further embodiment (FIG. 1b), an image of the mask can also be produced without a crossing point in this arrangement. If the image is reduced, the telecentricity is lost, although very small angles (approx. 6 mrad) on the wafer are still possible. In order to obtain a telecentric system again, a correction lens WL can be arranged according to the invention directly in front of the wafer (FIG. 8), through which the beams again strike the wafer almost axially parallel. According to the invention, this correction lens can be formed in that a single, essentially tubular electrode is positioned directly in front of the wafer and has slightly different potential than this. Together with the wafer, this electrode forms the so-called wafer lens WL, which again has a plate, namely the wafer itself, as an electrode and can therefore be operated as a diverging lens. This is the case if the potential at the electrode lying in front of the wafer is selected such that the ions are decelerated towards the wafer.

In a specific embodiment of the three-electrode lens with a grid, the distance between the mutually facing ends of the tubular electrodes R1, R2 is 135 mm. The electrode designed as a mask M is arranged at a distance of 90 mm from the outlet opening of the tube electrode R 1, the diameter of the outlet opening being 600 mm, whereas the diameter of the inlet opening of the tube electrode R2 is 300 mm. The outlet mouth of the tube electrode R2 is 675 mm away from the electrode designed as a mask M and the inlet mouth of the tube electrode R1 is 1350 mm in front of the mask forming the center electrode. In one The distance of 585 mm from the outlet mouth facing the mask M reduces the inside diameter of the tubular electrode R 1 from 600 to 300 mm.

In the embodiment variant shown in FIG. 6, the mask M forms the first electrode of the diverging gland G2. This is followed by the converging lens LI, followed by the second converging lens L2, which lies in front of the wafer plane in the beam path and is designed as an asymmetrical SINGLE lens. Since the potential of the last electrode of a lens is always the same as that of the first electrode of the subsequent lens, these two electrodes can each be physically formed as a single electrode, as can also be seen in FIG. 6. The entire ion-optical column between the mask and the wafer can therefore also be regarded as a single “multi-electrode lens”, in this case as a five-electrode lens, which has different refractive regions, namely G2, LI and L2. If necessary, for example in order to shorten the distance between the source and the wafer, an illumination lens can also be provided in front of the mask (one that is not shown in FIG. 6).

In contrast to the three-electrode lens described above, in which the mask M forms the central electrode, the field strength before and after the mask is different in the arrangement according to FIG. 6, so that there are flies' eyes lens effects caused by the Openings in the mask M. However, since the electric field strength at the level of the mask M, which faces the second electrode, is very small, and because the mask is the object to be imaged, the effect of these mini lenses is negligible. The low field strength is desirable because, due to the work that a film moving in an electric field has to do, any vibrations of the mask M are avoided. This allows the mask M to be changed and set up quickly. The arrangement according to FIG. 6 allows the image of the mask openings to be thrown onto the wafer W level with a telecentricity better than 2 mrad. This high telecentricity results in a depth of field of about 10 μm in the plane of the wafer W. In this embodiment, not only is the imaging scale and the resolution maintained even when the wafer position changes in the optical axis, but the particularly low distortion error can also be maintained. In the concept according to FIG. 6, the ion beam strikes a 60 × 60 mm ² mask structure field at angles that are less than 1.5 ° at the edge of the mask field. The mask openings can be realized in a 3 μm thick silicon foil with a slope of, for example, 3 °. In this way, line screens with an opening width of, for example, 0.45 μm can be realized in a 3 μm thick masking film without shadow errors due to the

Mask edges arise. The distance between wafer W and mask M is, for example, 2171.51 mm and the largest diameter is 1180 mm.

Regardless of the implementation, the devices according to the invention have the following properties: a) In the case of arrangements with crossover, the beam is generally given by the converging lens G3N (FIG. 1) or G2 which follows the mask M and / or contains it and LI (FIG. 2) a pillow-shaped residual distortion (area A) which, after the crossover, changes into a barrel-shaped residual distortion (area B).

FIGS. 1 a and 2 illustrate that the distortion of the converging lens L2 arranged in front of the wafer W gives the rays an additional deflection, which reduces the region B of barrel distortion and generates an area A 'of pillow distortion adjoining the barrel barrel distortion. According to FIG. 1b, in the case of the arrangement without crossover, the pillow-shaped distortion generated by the converging lens G3N is reduced by the lens L2 and finally converted into a barrel-shaped distortion (area B).

There is a level behind the collecting lens L2 arranged in front of the wafer W, in which the distortions caused by the lenses lying between the mask and the wafer are compensated for (level of minimal distortion). However, this only applies to the third-order distortion error; the much smaller fifth-order and higher-order distortion errors remain, so that the distortion does not disappear entirely, but only has a pronounced minimum. Using A WL lens (see below) can also help to further minimize distortion.

The dependence of the residual distortion on the deviation in the beam direction of the wafer position from the target position runs in the embodiment proposed here according to FIG. 6 with practically the same increase as that in the Europ. Patent application 93 890 058.6 proposed device in which two three-electrode grid lenses are used to image the openings of a mask (Fig. 7), the distortion minimum is even lower than in the above application. Together with the high telecentricity, this leads to a depth of field of approx. 10 μm, which is much better than the depth of field that can be achieved in a conventional image with light optics. For example, in the case of photolithography with high-energy UV light (“deep UV”) when producing structures of 0.2 μm, the depth of field is also in the range of only a few tenths of a μm. In order to be able to produce integrated circuits with this shallow depth of field, the surface of the wafer must be leveled in the light optics and the lithography may only take place in the top layer ("top surface imaging"), which is associated with high costs.

b) In the arrangements according to FIGS. 1b and 2, as can be seen schematically from FIGS. 3a and 3b, there is also a compensation for the chromatic error by the effects of the two converging lenses LI and L2 for a certain image plane E behind that the second converging lens L2 arranged on the wafer W. A beam (shown in broken lines in FIG. 3) with slightly smaller energy E ₀ - Δ E ₀ than the target energy E ₀ is deflected more strongly in the lens G3 or G2 + L1 containing the mask M than an beam with Target energy (shown in full lines) consequently hits and hits the lens L2 arranged in front of the wafer W in a larger axis than the target beam and is therefore deflected more strongly against the optical axis due to its lower energy, so that it is at a certain distance behind that in front of the wafer W arranged collecting lens L2 hits the target beam. The first order chromatic error disappears at this level E. There also remains a residual error, the 2nd order chromatic error, ie an error proportional to the square of the energy deviation from the target beam. In the arrangement according to FIG. 1 a, the wafer lens WL can be used to minimize the chromatic error.

c) The three relevant levels, namely the Gaussian image level of the mask M, the level of minimal distortion (FIGS. 1 and 2) and the level of minimal chromatic error (FIG. 3) will generally not coincide. By a suitable choice of the lenses and the position of the ion source Q on the ion-optical axis, it can be achieved that the three planes mentioned coincide while maintaining the approximately parallel beam path at the output of the lens L2. If the wafer W is then arranged where the three planes coincide, this has the effect that both the distortion and the chromatic error are minimized.

In all of the embodiments described here with cross-over (FIG. 1 a, FIG. 2, FIG. 6), the highest possible particle energy is achieved at cross-over, which has the advantage that the so-called stochastic space charge effects are minimized. The cross-over is in all cases in the field-free space in front of the lens L2.

Stochastic particle deflections are dealt with, for example, in the work "stochastic ray-deflections in focused particle beams" by A. Weidenhausen, Speer and H.Rose in Optik 69 (1985), pages 126 to 134. The stochastic space charge effect leads to resolution errors in the imaging, ie to a reduction in the maximum achievable resolution of the ion optics. The strength of this effect depends, among other things, on the energy at the cross-over, namely according to the above work according to δx _^ E ⁵ ' ⁴ where δx is the deterioration in resolution and

E is the energy of the ions at the crossover.

The following table shows the calculation results for the exemplary embodiment of an ion projection lithography system according to the invention as shown in FIG. TABLE I

Distance source-mask 1.6 m Distance mask-wafer 2.2 m image field size on the mask 60x60 mm - ** 'magnification (reduction) 3: 1 voltage ratio at G2 1.082 voltage ratio at LI 18.16 voltage ratio between the first and third electrodes of L2 0.75 voltage ratio between the first and second electrodes of L2 0.1825

Max. Distortion * <10 nm max. chromatic broadening 50 nm ** 20 nm *** max. Beam divergence according to the lens system <2 mrad

* Max. Deviation of any pixel from the ideal image

** for energy smearing of the ions at the ion source outlet E / E = 0.0003, where E is the energy of the ions at the outlet of the ion source. *** for ion sources in which the ions at the exit are smaller

Have energy smear, namely E E = 0.0001.

Under the above term voltage ratio between two electrodes of a lens with the

Potentials Uj and U2 should be understood as the following relationship:

where U _{0 is} the potential at which the kinetic energy of the charged particle would be zero.

Claims

 Claims: I. Particle, in particular ion-optical imaging system, preferably for lithography purposes, with a particle, in particular ion source for imaging a structure located on a mask film in the form of one or more transparent locations, in particular openings via at least two electrostatic devices arranged in the beam direction in front of a wafer Lenses on the wafer, one of the lenses being a so-called grating lens, which is formed by one or two tube electrodes (R1.R2) and a plate with openings arranged in the beam path, the plate being arranged normal to the optical axis (D), thereby characterized in that the plate is formed by the masking film (M) which forms the middle or the first electrode of the grating lens in the beam direction. 2. Imaging system according to claim 1, characterized in that there is a further lens in the beam path between the mask and the wafer, which consists of a substantially tubular electrode and a plate, the plate being formed by the wafer, and the tubular electrode immediately before the wafer is positioned. 3. Imaging system according to claim 1 or 2, characterized in that the mask is the central electrode of a three-electrode grating lens, and the second lens between the mask and the wafer is designed as an immersion lens. 4. Imaging system according spoke 1 or 2, characterized in that the mask is the central electrode of a three-electrode grating lens, and the second lens between the mask and the wafer is designed as an asymmetrical single lens. 5. Imaging system according to claim 1, characterized in that the mask is the first electrode of a two-electrode grating lens, a second, subsequent lens as a two-electrode immersion lens and a third and last lens between the mask and the wafer is designed as an asymmetrical single lens. 6. Imaging system according to one of claims 1 to 5, characterized in that the crossing point of the ions is in the field-free space. 7. Imaging system according to claim 6, characterized in that the energy of the ions in the region of the crossing point is higher than on the wafer. 8. Imaging system according to one of claims 1 to 4, characterized in that the beam path between the mask and the wafer has no crossing point.